A1364
Title: A new multivariate two-sample rank test based on interpoint distances
Authors: Hikaru Yamaguchi - Tokyo University of Science (Japan) [presenting]
Hidetoshi Murakami - Tokyo University of Science (Japan)
Abstract: Testing general multivariate two-sample problems is important in many applications. Interpoint distance-based tests are well known to be effective for general alternatives, including simultaneous comparisons of means and variability, or comparisons of the shapes of the distributions. The aim is to propose a new test statistic based on the ranks of interpoint distances to address the two-sample problem for multivariate populations. The proposed test statistic is constructed from the well-known energy distance, and its critical values are approximated using a permutation procedure. Some theoretical properties of the proposed test statistic are also presented. Numerical simulations are conducted to examine the power and robustness of the test under various settings, including high-dimensional data and small sample sizes.
